Re: Issue 155: WTP/AC ACL Synchronization Issue
From: Nathan J. Williams (nathan.williamsthingmagic.com)
Date: Tue, 5 Aug 2008 15:35:03 -0700 (PDT)
"Pat Calhoun (pacalhou)" <pcalhoun [at] cisco.com> writes:

> Second, the way it works is that the AC *always* sends an updated list
> to the WTP at reboot. However, some WTPs *could* conceptually operate
> without an AC (such as the case where the WAN link between the WTP and
> the AC is down). Obviously, such WTPs could only operate this way when
> configured for Local mode. Split mode requires an AC.
>
> So the WTP stores the ACL, and can use it even if it reboots and it is
> unable to connect to the AC. However, it always gets an updated list
> from the AC when it does connect.
>
> So... Given that we've never discussed the concept of a WTP operating
> without an AC, does it really make sense to leave this functionality in
> the spec? I think we should pull out the Add/Delete Static message
> elements, and deal with this in a future rev of the spec if the need
> arises (because it will likely require more than just some verbage here
> and there).
>
> Thoughts?

Deleting the Add/Delete Static message elements sounds good to
me. Should we spell out when the WTP's list is reset to zero? At the
time of association with an AC seems right, otherwise things like
Primary Discovery could re-introduce the problem of losing sync of the
ACL.

        - Nathan

Results generated by Tiger Technologies using MHonArc.